A concussion is a hit to the blood supply.

The flow drops, and the regulator fails.

A concussion is not only a metabolic event; it is a vascular one. Within 24 to 48 hours of a sport-related concussion, arterial-spin-labeling MRI picks up reduced cerebral blood flow against matched controls, across parietal, frontal and cingulate regions.¹ The brain’s own pressure regulator, dynamic cerebral autoregulation, falters too: in athletes tested with squat-stand manoeuvres, autoregulatory phase dropped about 23 percent at 72 hours and was still down roughly 18 percent at two weeks, recovering only by one month, while players were cleared to return at a median of 14 days.² That gap is the point. The plumbing is still impaired after the scoreboard says you are fine. Underlying it is a known failure mode: after head trauma, autoregulation can break so that flow passively follows blood pressure, leaving regions exposed to hypoperfusion and secondary ischemic injury when pressure dips.³ And slower recovery of flow tracks slower clinical recovery: in athletes imaged serially, depressed insular blood flow persisted at one month in the slower-to-recover, correlating with worse mood.⁴

There is no single concussion scan, and that is the honest part.

It would be tidy if a brain scan could read out a concussion. It cannot, and the literature says so plainly. The direction of the flow change is not uniform: acutely, athletes with cognitive-predominant symptoms show lower frontal and subcortical flow, while those with somatic-predominant symptoms show higher posterior flow, which is exactly why a group average can look unchanged even when individuals are abnormal.⁵ Hypo- and hyperperfusion can coexist in the same brain.⁶ A systematic review of 23 ASL studies found regional abnormalities in most, with more decreases than increases, but called for larger cohorts and stopped short of a diagnostic signature.⁷ At two weeks in children, global perfusion did not separate the still-symptomatic from the recovered.⁸ The sport-concussion consensus is explicit: perfusion and functional imaging are research findings, not clinical diagnostic criteria.⁹ The vascular story is real at the group level and genuinely useful for mechanism. It is not a test you should ask for to tell you whether you, personally, are concussed.

What helps, what does not, and the one thing that kills.

For decades the advice was to lie in a dark room until you felt normal. That was wrong. A randomized trial found five days of strict rest produced more symptoms over ten days than the usual 1-2 days of rest then gradual return, and a meta-analysis pooling 4,329 patients found prescribed rest slightly worsened symptoms with no recovery benefit.¹⁰¹¹ What does work is the opposite, dosed carefully. The Buffalo Concussion Treadmill Test finds the heart rate at which symptoms flare, and exercise is prescribed just below it; a landmark RCT in adolescents had the aerobic group recover in a median 13 days versus 17 on stretching, and a larger multicentre trial roughly halved the risk of persistent symptoms.¹²¹³ The mechanism is vascular: concussion leaves ventilation too low for effort, raising CO2 and driving flow up out of proportion to demand, and graded sub-threshold work retrains that response toward normal CO2 sensitivity and exercise blood flow.¹⁴¹⁵ Two firm warnings. Stay below your threshold, not above it; if exertion crashes you, this is sub-threshold dosing, not push-through. And return to play is deliberately conservative because of second-impact syndrome: a second hit before the first heals is thought to crash autoregulation into catastrophic swelling, with collapse in minutes, overwhelmingly in the young, very rare but near-uniformly devastating, which is why the protocol is graduated, at least 24 hours per step, and return-to-learn comes before the final contact stages.¹⁶¹⁷¹⁸

Cleared is not the same as healed.

The years-later damage collects around the same small vessels.

Repeated head impacts leave a long tail, and its anatomy is strikingly vascular. The defining lesion of CTE is hyperphosphorylated tau clustered around small blood vessels at the depths of cortical sulci, and unlike Alzheimer’s diffuse spread, CTE tau characteristically begins in the perivascular space and works outward.¹⁹²⁰ The blood-brain barrier appears to fail at exactly those vessels: tight-junction proteins break down where perivascular tau deposits, leakage markers colocalize with the tau and scale with years of contact-sport exposure, and a reactive glial halo concentrates within tens of micrometres of the lesion’s central vessel, hinting that barrier injury may be upstream, not just downstream, of the tau, though that remains a hypothesis.²¹²²²³ In living retired athletes, barrier leakage persists years after retirement and tracks worse cognition, and former players show widespread chronic hypoperfusion.²⁴²⁵ The risk is real but dose-dependent, not universal: dementia risk is markedly elevated in professional soccer and boxing, against a general-population benchmark where hospitalized head trauma roughly doubles later dementia, while high-school American football in one pooled estimate was not elevated.²⁶²⁷

The same upright-flow drop shows up after a head injury, see standing intolerance. Post-traumatic headache usually takes a migraine form. The rehab is the sub-threshold version of what helps, and repeated impacts feed the aging brain.